On Tue, 2006-12-05 at 10:02 +0100, Michele Baldessari wrote: > The previous version had "inlen != 2" test which doesn't really make > much sense since we are checking the two next characters, so we need to > guarantee their presence.
Correct, I made a typo there. > I also added a second check for trailing of > single '\r' and '\n' only, which was why it was barfing on my testcase > with sasl-simple-client. Good to see that you came to the same conclusion independently! :) My patch has all the checks in one if clause, so it doesn't duplicate the check for saw_equal. It also catches the case where inlen is not 1 or 2 after the > 0 check -- it could be 3 after the decoding while loop (the while condition is inlen > 3) in which case it's either a protocol / alignment error or a partial base64 block. Perhaps it's also a little bit easier to understand since all conditions are positive. Other than these differences, your patch seems to be equivalent. Yours is prettier, though, but I'm afraid that's secondary. :) If you see any mistakes in my patch, then please yell! An upload will be made within a day or two. Cheers, -- Fabian Fagerholm <[EMAIL PROTECTED]>
signature.asc
Description: This is a digitally signed message part